REFINED STRUCTURE OF A BENCE-JONES PROTEIN
`
`Sipos, T., and Merkel, J. R. (1970), Biochemistry 9. 2766.
`Smith, R. L., and Shaw, E. (1969), J. Biol. Chem. 244,
`4704.
`Solomon, I. ( 1955), Phys. Rev. 99, 559.
`Stroud, R. M., Kay, L. M., and Dickerson, R. E. (1971),
`Cold Springs Harbor Symp. Quant. Biol. 36, 125.
`Stroud, R. M., Kay, L. M., and Dickerson, R. E. (1974), J.
`Mo/. Biol. 83, 185.
`Titani, K., Ericsson, L. H., Neurath, H., and Walsh, K. A.
`(1975), Biochemistry 14, 1358.
`Trowbridge, C. G., Krehbiel, A., and Laskowski, M.
`( 1963), Biochemistry 2. 843.
`
`Valenzuela, P., and Bender, M. L. (1969), Proc. Natl.
`Acad. Sci. U.S.A. 63, 1214.
`Valenzuela, P., and Bender, M. L. (1970), Biochemistry 9,
`2440.
`Van Geet, A. L., and Hume, D. N. (1965), Anal. Chem. 37,
`979.
`Villanueva, G. B., and Herskovits, T. T. (1971), Biochemis(cid:173)
`try JO, 4589.
`Walsh, K., and Neurath, H. (1964), Proc. Natl. Acad. Sci.
`U.S.A. 52, 884.
`Yguerabide, J., Epstein, H. F., and Stryer, L. (1970), J.
`Mo/. Biol. 51, 573.
`
`The Molecular Structure of a Dimer Composed of the
`Variable Portions of the Bence-Jones Protein REI
`Refined at 2.0-A Resolution t
`
`Otto Epp,* Eaton E. Lattman,t Marianne Schiffer,§ Robert Huber, and Walter Palm#
`
`ABSTRACT: The structure of the variable portions of a
`K-type Bence-Jones protein REI forming a dimer has been
`determined by X-ray diffraction to a resolution of 2.0 A.
`The structure has been refined using a constrained crystal(cid:173)
`lographic refinement procedure. The final R value is 0.24
`for 15,000 significantly measured reflections; the estimated
`standard deviation of atomic positions is 0.09 A. A more
`objective assessment of the error in the atomic positions is
`possible by comparing the two independently refined mono(cid:173)
`mers. The mean deviation of main-chain atoms of the two
`chains in internal segments is 0.22 A, of main-chain dihe(cid:173)
`dral angles 6.3° for these segments. The unrefined molecu(cid:173)
`lar structure of the VREI dimer has been published (Epp,
`0., Colman, P., Fehlhammer, H., Bode, W., Schiffer, M.,
`Huber, R., and Palm, W. (1974), Eur. J. Biochem. 45,
`513). Now a detailed analysis is presented in terms of hy(cid:173)
`drogen bonds and conformational angles. Secondary struc-
`
`tural elements (antiparallel (3 structure, reverse turns) are
`defined. A more precise atomic arrangement of the amino
`acid residues forming the contact region and the hapten
`binding site is given as well as the localization of solvent
`molecules. Two cis-prolines (Pro-8 and Pro-95) were de(cid:173)
`tected. The intrachain disulfide bridge (Cys-23-Cys-88) oc(cid:173)
`curs statistically in two alternative conformations. The
`structure suggests reasons for strong conservation of several
`amino acid residues. The knowledge of the refined molecu(cid:173)
`lar structure enables crystal structure analyses of related
`molecules to be made by Patterson search techniques. The
`calculated phases based on the refined structure are much
`improved compared to isomorphous phases. Therefore the
`effects of hapten binding on the molecular structure can be
`analyzed by the difference Fourier technique with more re(cid:173)
`liability. Hapten binding studies have been started.
`
`lmmunoglobulins are proteins with specific antibody ac(cid:173)
`tivity. There exist several classes. The IgG class of immuno(cid:173)
`globulins is composed of two light and two heavy chains.
`The Bence-Jones proteins excreted by patients with multi(cid:173)
`ple myeloma into the urine have been shown to be free light
`chains. The Bence-Jones protein REI is a human immuno(cid:173)
`globulin light chain of K type. The purification, crystalliza-
`
`t From the Max-Planck-lnstitut fiir Biochemie, 8033 Martinsried
`bei Miinchen, West Germany, and Physikalisch-Chemisches Institut
`der Technischen Universint, Miinchen. Received May 30, 1975. The
`financial assistance of the Deutsche Forschungsgemeinschaft and Son(cid:173)
`derforschungsbereich 51 is gratefully acknowledged.
`I Present address: Rosenstiel Institute, Brandeis University, Wal(cid:173)
`tham, Massachusetts 02154.
`§Was on leave from: Division of Biological and Medical Research,
`Argonne National Laboratory, Argonne, Illinois 60439.
`# Present address: Institut fiir Medizinische Biochemie der Univ(cid:173)
`ersitat Graz, Austria.
`
`tion, and sequence analysis has been described (Palm, 1970;
`Palm and Hilschmann, 1973, 1975; Palm, 1974). The crys(cid:173)
`tal structure of a dimer composed of the variable portions of
`this Bence-Jones protein at a resolution of 2.8 A was re(cid:173)
`ported (Epp et al., 1974). Data to a resolution of 2.0 A have
`now been collected and the structure has been refined by
`constrained crystallographic refinement. The aim was to get
`a detailed insight into the conformation of this molecule
`(main chain, side chains, and bound solvent) and to obtain a
`model sufficiently accurate for its use in Patterson search
`techniques to determine the crystal structures of related
`molecules (Fehlhammer et al., 1975). As refined phases are
`considerably better than isomorphous phases (Watenpaugh
`et al., 1973; Deisenhofer and Steigemann, 1975; Huber et
`al., 197 4 ), the quality of difference Fourier maps will be
`much improved; this will make it possible to determine the
`structure bound haptens and the subtle structural changes
`which might occur upon binding.
`
`BIOCHEMISTRY, VOL. 14, NO. 22, 1975 4943
`
`1 of 10
`
`BI Exhibit 1113
`
`

`

`EPP ET AL.
`
`(3) Structure-Factor Calculation
`
`(W. Steigemann and T. A. Jones) atomic scattering factors: Forsytli
`and Wells constants (1959), an overall temperature factor was used
`throughout.
`
`(4) R Value Calculation
`
`IFc II
`
`R is defined as
`
`2: llF0 I
`L:tF0 1
`IF0 i, observed structure--factor amplitude
`iF c i, calculated structure -factor amplitude
`For this calculation, as well as for the Fourier calculations, reflec(cid:173)
`tions of extremely bad correlation were excluded. The condition for
`exclusion was:
`
`tFcil
`
`211F0 1
`> 1.2
`l!F0 I + iFc II
`About 250 reflections were excluded. The innermost reflections to
`6.8-A resolution were omitted from all calculations (941).
`
`(5) Fourier Synthesis
`
`Grid: 1.1 Ax 1.1 A X 1.0 A at the beginning to a resolution of 2.4
`A, later 0.8 A X 0.8 A X 0.8 A. During the course of the refinement
`the Fourier coefficients were mostly of the type (2 IF0 I -
`!Fe I)
`exp°'c· Also some Fourier syntheses with coefficients (3 IF 0 I -
`2 IFc I) expo:c were used. Difference Fourier maps were calculated
`!Fell expo:c'°'c• calculated phase.
`with coefficients (F0 1 -
`
`(6) Computer Time on a Siemens 4004/150 (Cycle Time 0.75 µsec)
`
`Structure-factor calculation for 15,000
`reflections and 1630 atoms
`Fourier synthesis 1.62 X 10 5 grid points
`Real-space refinement for both chains in
`the asymmetric unit
`
`13,500 sec
`
`4000 sec
`17 ,000 sec
`
`Table I: Options and Specifications of Various Programs.
`
`(1) Model Building (Diamond, 1966)
`
`Probe
`
`1
`2
`3
`4
`5
`
`Probe
`Length
`oa
`2
`5
`6
`6
`
`I
`I
`2
`5
`6
`
`0.5
`0.5
`0.1
`0.1
`0.0
`
`1.()ll
`0.5
`0.1
`0.0
`0.0
`
`Variation of
`Filter Constants
`----------~- Dihedral
`c,
`C,·
`Angles
`10-2a
`10-·
`10-•
`10- 2
`10-2
`
`10-•
`10-•
`10-•
`10-•
`10-•
`
`¢,>i;,x
`¢,y,X,7
`¢,y,X,T
`¢,t/J,X,T
`¢,y,X,7
`
`Eigenshifts are permitted if either?-..:> C,e' or Jc :> C2?.max•
`where A are eigenvalues of the normal matrix. e 2 is the residual.
`Variation of the folds of pralines, x' of arginine, and w was not
`permitted. The angular value of T (N-Coi-Cl was fixed to a
`value of 109.65° including model-building procedure 7. In
`model-building 7 cis-prolines were introduced (Huber and
`Steigemann, 197 4 ). ¢,y are main-chain dihedral angles, x are
`side-chain dihedral angles. The angular values of main chain
`and side chains were used and included wherever they were
`known, even if only crudely.
`
`(2) Real-Space Refinement (Diamond, 1971, 1974)
`
`Zone length
`Margin width
`FL'\ed radius of all atoms (A)
`Relative weights of C:N:O:S
`Relative softness of dihedral angles
`y,¢,x1-x•
`x'
`6 1,e ',e" (pro line)
`e 1,e' (cysteine)
`T (N-Cu-C)
`w (torsion angle Ci-1°'-Ci-1-Ni--C{')
`
`7
`6;8b
`1.55; l.SQC
`6:7:8:16
`
`100
`1
`0.1
`0
`0.1
`0.ld
`
`Refinement of Scale Factor (K) and Background Level (d) Only
`
`Filter ratio Amin/Amax for K, d refinement
`Filter ratio for rotational refinement Amin/Amax
`Isomorphous map
`from difference Fourier map 2 on
`from difference Fourier map 4 on
`from difference Fourier map 6 on
`The value of 0.00004 resulted in a proportion
`of ~hifts applied of about
`Filter ratio for translational refinement
`Amin/Amax
`The value of 0.01 resulted in a proportion of
`shifts applied of about
`
`0.01
`
`0.005
`0.0005
`0.00007
`0.00004
`
`70%
`
`0.01
`
`100%
`
`a These values were used in model-building procedures 8 and 9 for probe length and filter constants. Also the variation of the angular value
`of T was allowed, but not for Gly, Ser, and Thr. b After introduction of w. c New value estimated with the data at 2.1-A resolution. d w has
`been introduced after difference Fourier map 4 (data to 2.3-A resolution,R factor value 0.29,B = 23 A 2
`).
`
`_______________ :__:__ ___________________________________ _
`
`Experimental Procedure
`The V REI dimer crystallizes in the space group P6 1• The
`hexagonal unit cell parameters are a = b = 75.8 A, c =
`98,2 A, 'Y = 120°. The asymmetric unit contains one dimer
`molecule. Intensity data were collected to 2.0-A resolution
`on a modified Siemens AED diffractometer using 8-28
`scan, with focus-to-crystal and crystal-to-detector distances
`of 30 cm each. The intensity profile of each reflection was
`scanned twice in steps of 1/100° ( 44 steps over the whole
`reflection). The counting time per step was set inversely
`proportional to the intensity at the peak of the reflection;
`the upper limit was set at 2.4 sec/step for reflections to a
`resolution of 2.5 A. The reflections from 2.5- to 2.0-A reso(cid:173)
`lution were measured with 6 sec/step. Background was
`counted on both sides. The whole data set had been collect(cid:173)
`ed by diffractometers. The reflections were collected in
`shells of sin 8/>... Film data obtained from screenless preces(cid:173)
`sion photographs which were evaluated using the method of
`
`Schwager et al. (Schwager et al., 1975) were included and
`used for scaling purposes. The data were corrected for ab(cid:173)
`sorption by an empirical method (Huber and Kopfmann,
`1969). The complete intensity data set included 42,464
`measurements (of which 5454 are film data), which were
`merged to 14,993 independent reflections (0.69 of the possi(cid:173)
`ble reflections to a resolution of 2.0 A), All reflections to a
`resolution of 2.5 A were used ( 10381 ), except the innermost
`reflections to a resolution of 6.8 A (941) which are strongly
`influenced by the solvent continuum. To a resolution of 2.5
`A, 0.99 of the possible reflections is measured. From 2.5- to
`2.0-A resolution, only the reflections which could be ob(cid:173)
`served above the 20' significance level determined from
`counting statistics (3671) were used (e.g., 0.34 of the possi(cid:173)
`ble reflections in that range).
`The neglect of the innermost reflections to a resolution of
`6.8 A does not influence Diamond's real-space refinement
`procedure. In the refinement of PTI (Deisenhofer and
`
`4944 BI 0 CHEMISTRY, Y 0 L. 1 4, N 0. 2 2,
`
`I 9 7 5
`
`2 of 10
`
`BI Exhibit 1113
`
`

`

`REFINED STRUCTURE OF A BENCE-JONES PROTEIN
`
`Table II: Specifications of Difference Fourier Maps.
`
`Corrections
`
`R = (~ * ((/)h -
`
`Steigemann, 1975) the neglect of the innermost reflections
`improved the refinement of the positions of solvent m.ole(cid:173)
`cules. We omitted these reflections from all calculations
`and did not consider them in the interpretation of the elec(cid:173)
`tron density map. The R value for all measurements, de(cid:173)
`fined as
`
`lhj)
`
`2)1 12
`
`2 /~ Nh(I)h
`is 0.05 ( (/) h is the average intensity of the Nh measure(cid:173)
`ments, /hi are the individual measurements of a reflection
`h). The Rsym values for individual crystals lie between
`0.022 and 0.065; the average Rsym is 0.04.
`The crystal structure of REI has been refined by a con(cid:173)
`strained crystallographic refinement described in the refine(cid:173)
`ment of the crystal structure of the bovine pancreatic tryp(cid:173)
`sin inhibitor (PTI) (Deisenhofer and Steigemann, 1975)
`and in the refinement of the structure of the complex be(cid:173)
`tween bovine trypsin and PTI (Huber et al., 1974). This
`procedure involves cycles consisting of phase calculation
`using the current atomic model, Fourier synthesis using
`these calculated phases and the observed structure-factor
`amplitudes, and Diamond's real-space refinement (Dia(cid:173)
`mond, 1971, 197 4 ). At various stages (stagnation of further
`refinement), difference Fourier syntheses are calculated to
`detect and correct gross errors in the model (such as incor(cid:173)
`rect orientation of main-chain amides or side chains) and to
`localize solvent molecules. The stagnation of the refinement
`is reached, if the R factor value (definition see Table I)
`does not decrease further. Incorrectly oriented main-chain
`amides are rotated by reading the correct position of the
`carbonyl oxygen from the difference map and subjecting
`the whole chain to a model-building procedure (Diamond,
`1966). Side-chain orientations are corrected by rotating
`around the appropriate dihedral angles. Some characteris(cid:173)
`tics of the model-building procedure and the real-space re(cid:173)
`finement procedure are outlined in Table I. During the re(cid:173)
`finement 8 model-building procedures and 11 difference
`Fourier syntheses were calculated. The Fourier syntheses in
`the automatic cycles between these difference Fourier maps
`were mostly of the type {2j F 01 - IF J) exp ac; a few (31 F J -
`2J F J) exp ac were also used. Such syntheses increase the
`density gradients at the atomic positions and the speed of
`convergence. Table II shows the statistics of the course of
`the real-space refinement procedure.
`
`Results and Discussion
`Description of the Refinement. The starting model had
`been obtained through extensive real-space refinement of
`the model into the isomorphous Fourier map at 2.8-A reso(cid:173)
`lution. The above electron density map had first been aver(cid:173)
`aged over the two independent molecules {Epp et al., 1974).
`The starting R factor (defined in Table I, 4) was 0.48. Dur(cid:173)
`ing the subsequent course of the refinement the two mole(cid:173)
`cules in the asymmetric unit were refined independently.
`The coordinates of the second molecule were obtained by
`applying the known local symmetry. After several cycles the
`R factor decreased to a value of 0.39. At this stage, the first
`difference Fourier map was calculated. The coordinates
`were plotted onto the Fourier map to check the progress.
`Misplaced side chains and several incorrectly oriented
`main-chain amides were detected (Table II). Between suc(cid:173)
`ceeding difference Fourier maps three to five automatic re(cid:173)
`finement cycles were performed, depending on the progress
`of the refinement. The refinement was started with reflec-
`
`in Each Map
`Re so- Overall Atom
`Side Chains
`Main.Chain Amides lution Temp Fact. Radius
`Difference
`(A 2)
`Map No. R Value Chain 1 Chain 2
`<r>(A)
`(A)
`7
`1
`10
`0.390
`6
`4
`0.350
`2
`2
`0.304
`3
`3
`11 solvent molecules
`2
`4
`26 solvent molecules
`3
`4
`40 solvent molecules
`1
`1
`39 solvent molecules
`introduction of cis-
`pralines (Pro-8,
`Pro-95)
`1
`42 solvent molecules
`1
`1
`36 solvent molecules
`
`4
`
`5
`
`6
`
`7
`
`8
`
`0.294
`
`0.282
`
`0.270
`
`0.264
`
`0.250
`
`9
`
`10
`
`11
`
`0.241
`
`0.241
`
`0.241
`
`41 solvent molecules
`
`55 solvent molecules
`
`5 3 solvent molecules
`
`2.5
`2.4
`2.4
`
`2.3
`
`2.3
`
`2.2
`
`2.2
`
`2.2
`
`2.1
`
`2.1
`
`2.0
`
`28
`25
`23
`
`23
`
`23
`
`23
`
`23
`
`23
`
`23
`
`23
`
`21
`
`1.55
`1.55
`1.55
`
`1.55
`
`1.55
`
`1.55
`
`1.55
`
`1.55
`
`1.55
`
`1.50
`
`1.50
`
`tion data to 2.5-A resolution. During the course of the re(cid:173)
`finement further data were included to 2.0-A resolution. An
`overall temperature factor was used, which was recalculat(cid:173)
`ed several times by comparing IF J and IF J and the temper(cid:173)
`ature factor changed if necessary. An average atomic radius
`of 1.55 A was used for all atoms throughout the refinement
`until difference Fourier map 9. Afterwards a new value of
`1.50 A was estimated by a trial calculation of atomic radii
`refinement in the real-space refinement step. This value is
`directly related to and consistent with the observed overall
`temperature factor of 21 A2• By the inspection of the differ(cid:173)
`ence Fourier maps, solvent molecules were detected and
`used in the phase calculations. The course of the difference
`maps is shown in Table II. The starting model had to be
`corrected in several segments. A number of main-chain am(cid:173)
`ides and side chains had to be rotated. For the correction of
`main-chain amides, the presence of two identical molecules
`in the asymmetric unit provided a very useful cross-check.
`In difference Fourier map 6, the segments around Pro-8
`and Pro-95 could be corrected. The local distribution of
`maxima and minima in these two regions was very similar
`and consistent in the two independent molecules, and it sug(cid:173)
`gested the presence of cis-peptide groups (Huber and
`Steigemann, 1974). After the introduction of these cis-pro(cid:173)
`lines, the refinement proceeded and stopped finally at an R
`factor value of 0.24. At this stage, 53 solvent molecules had
`been identified.
`Description of the Electron Density Map. Figure la and
`b is stereo pictures of the electron density and the model fit
`at two amino acid residues (G ln-3 7 and Tyr-71) 1 in order to
`
`1 Amino acid residue numbers are those of the V REI sequence. The
`nomenclature recommended by IUPAC- IUB (1970) is used in this
`paper with additional definitions as given by Diamond ( 1966, 1971 ,
`1974) . The coordinates of the VR E I dimer are available upon request.
`They are also in the Protein Data Bank of Brookhaven National Labo(cid:173)
`ratory. Coordinates as well as stereo drawings are contained in R.
`Feldmann's Global Atlas of Protein Structure on Microfiche.
`
`BI 0 CHEMISTRY, V 0 L.
`
`I 4, N 0. 2 2, I 9 7 5 4945
`
`3 of 10
`
`BI Exhibit 1113
`
`

`

`EPP ET AL.
`
`R
`
`\
`
`FIGURE 1: Electron density and model fit at two amino acid residues (top, Gln-37; bottom, Tyr-71 ). Contours in these figures from 0.3 e/ AJ in
`steps of0.3 e/A 3 .
`
`demonstrate the quality of the final Fourier map. A typical
`electron density for carbon is 0.8 e/ A3 and for oxygen 0.9
`e/A3.
`The final difference Fourier map is clear and suggests
`that the molecular structure is correctly interpreted. There
`is still some positive and negative residual density. Because
`of the restricted rotations about the S-S bond, cystine can
`exist in two mirror-image forms (Beychok, 1967). In the
`VREl 2 monomer there is one intrachain disulfide bridge
`(Cys-23-Cys-88). Both conformations, differing predomi(cid:173)
`nantly in the position of 23 SY, occur statistically in about
`the same proportion in each monomer. This is shown in Fig(cid:173)
`ure 2a and b where sections 27-34 of difference Fourier
`maps (F0
`- Fe) based on both conformations are drawn for
`monomer 2. Figure 2a shows the final difference map.
`Figure 2b shows a difference map based on the alterna(cid:173)
`tive cystine conformation in the molecule. The difference
`between the two configurations is essentially expressed in
`the value of the side-chain dihedral angle x3 (rotation about
`the S-S bond). It differs by 162°. The final difference Fou(cid:173)
`rier map also shows two statistical positions for the side
`chain of Gln-100.
`Besides these there are about 40 negative peaks with
`height -0.20 e/ 1\3, 20 with height -0.25 e/ A3, and a few
`with height -0.30 e/A3. About 15 residual density peaks
`are positive, the highest is +0.25 e/ A3 . The positive peaks
`are near atomic positions, the negative features occur main(cid:173)
`ly at carbonyl oxygens, at the badly defined region of the
`N-terminus, at Gly-41, and at C' and NI of lysine side
`chains. The negative residual density at carbonyl oxygens
`has also been observed in the PTI-trypsin complex and
`suggests that main-chain vibration affects predominantly
`
`2 Abbreviations used are: Fab', antigen-binding fragment of immu(cid:173)
`noglobulins; V REI, variable part of the Bence-Jones protein REI; V L.
`variable part of light chain; LI, L2, and L3 are the first, second, and
`third hypervariable regions of light chains; K, ,\, the two major types of
`light chains differentiated by their C-terminal amino acid sequences:
`Dnp, 2,4-dinitrophenyl.
`
`the carbonyl oxygens. Six of the 53 water molecules found
`at the end of the refinement, and included in the calcula(cid:173)
`tions with full occupancy and a temperature factor of 21
`A 2 , lie in negative density of -0.25 e/ A 3 .
`Accuracy of the Structure; Comparison of the Two Mo(cid:173)
`nomers. A comparison of the dihedral angles of the two
`main chains shows that only at the N- and C-termini, which
`are badly defined, at some Gly residues (Gly-57, Gly-64),
`and in some external loops (12-18, 26-34 first hypervaria(cid:173)
`ble region, 40-44, 78-81, 93-97 part of the third hypervar(cid:173)
`iable region) the differences between the two chains are
`considerably larger than the mean deviation (/l</>,i/;) of
`9.6°. There are particularly large discrepancies in</> and if;
`at Ser-26, Gln-27, and Asp-28. We observed no significant
`deviation in dihedral angles of residues forming the mono(cid:173)
`mer-monomer contact across the local diad (see Figure 5).
`The segments deviating significantly from the local symme(cid:173)
`try face the solvent, and the reasons for these structural dif(cid:173)
`ferences are unclear. As the difference map is featureless in
`these regions (Gln-27 is visible only to CY in both chains)
`these structures appear to represent real alternative confor(cid:173)
`mations. The mean deviation excluding these segments is
`(t!>.f/>,i/;): 6.3°. The same result is reflected by a comparison
`of the main-chain atoms (N, C", Ci3, C, 0) of internal seg(cid:173)
`ments of both monomers (segments 4-10, 19-25, 35-38,
`45-49, 62-77, 84-89, and 97-102). The mean deviation (r)
`is 0.22 A compared with 0.47 A for all atoms (the mean de(cid:173)
`viation of the internal segments for PTI and PTI complexed
`with trypsin is 0.25 A, Huber et al., 1974). The average
`shift of main-chain atoms from the starting model (identi(cid:173)
`cal monomers) to the final model was 0.79 A neglecting N(cid:173)
`and C-terminal segments. A comparison of side-chain dihe(cid:173)
`dral angles shows larger differences, especially at residues
`Ser, Thr, Gin, and Lys.
`An objective assessment of the error of 0.2 A, in the
`atomic coordinates of the final model, is possible by com(cid:173)
`paring the two chains forming the dimer. This estimate is
`an upper limit because there may be small structural alter-
`
`4946 BI 0 CHEMISTRY, V 0 L. I 4, N 0. 2 2, I 9 7 5
`
`4 of 10
`
`BI Exhibit 1113
`
`

`

`REFINED STRUCTURE OF A BENCE-JONES PROTEIN
`
`a
`
`'.,,/~:5CZ2
`
`'· ',' JJCDJ
`
`D
`
`llN
`
`0
`ll C!
`
`21 C01 ,o,. .. -
`
`O I NOE
`
`.
`:_-:.·-; ... ·
`
`,
`
`27
`
`~OG1
`
`'
`? 'CD
`'
`(J b •NOE
`
`'
`
`28
`
`tJ
`
`oJ3 COi
`
`', . •
`
`71 CB °'
`~:_',
`
`71 CG 'a
`
`USG
`
`0 99 0
`
`/.'--·
`
`JO
`
`lc3co1
`I o6NOES_-,-,
`( _ .... __ :. ..... _ )
`
`0
`
`6 UCB
`
`27
`
`I L .: ,.
`
`29
`
`033 CD1
`0
`
`!O
`
`71 CA o
`
`0
`71 N
`
`0
`'CG
`
`0
`
`99 N
`9 8 c
`
`/
`
`!1
`
`l3 CA
`0
`
`I
`
`~ I
`,';:~ __ @\X/ I
`
`71 CA o
`
`}
`
`I
`71 CD1 0 I
`--
`I
`/ :
`
`2! CA
`0
`
`0 ©'
`
`I
`
`I
`
`~
`
`I
`
`1,:
`
`I
`
`!!
`
`J 1
`
`I
`~
`
`!3
`
`L-~~~
`
`!2
`
`J4
`
`J2
`
`22 c
`22~ 0
`
`2J N
`
`71 N
`
`o 6 CB
`0
`
`6 c
`
`r
`
`• ,6 CA
`
`I
`
`'
`
`'•
`!.
`
`'-
`
`.,.-ell 0
`23 c
`
`q,
`d I
`I
`?so,-,
`'
`
`' CE
`
`'
`,.,-;
`I ; ... > I
`
`FIGURE 2: Sections 27-34 of difference Fourier maps (F0 - Fe) based on one of both possible cystine (Cys-23-Cys-88) conformations, respectively
`(shown is monomer 2), Contours from 0.10 e/A 3 in Steps of0,05 e/A3: (-)positive, (- - -) negative residual densities. (a) Final difference map; (b)
`an intermediate state of the refinement. This explains the high residual density below Met-4 SD.
`
`ations on dimer formation,
`An estimate of the accuracy of the final atomic positions
`can also be deduced by Cruickshank's formulas using the
`residual and the curvature of the electron density at the
`atomic positions (Cruickshank, 1949), The mean height of
`carbonyl oxygens in the final Fourier map is 0,9 e/ A3. This
`yields a curvature of -1.32 e/ A 5 (Stout' and Jensen, 1968;
`Watenpaugh et al., l 973). The resulting standard deviation
`is o-(x) = 0.09 A. This is comparable to the values obtained
`for PTI-trypsin complex (Huber et al., 1974).
`Description of the Monomer. The VREI monomer has a
`sandwich-like structure. The polypeptide chain can be di(cid:173)
`vided into nine segments, which form two halves partly con(cid:173)
`sisting of anti-parallel f3 structure. These segments are con(cid:173)
`nected by reverse turns. The two sheets cover a hydrophobic
`interior containing several invariant or almost completely
`conserved amino acid residues in K-light chains (Epp et al.,
`1974). The upper part consists of five strands, the bottom
`
`part of three strands. The N-terminal strand adds to the
`upper sheet in a parallel and to the lower sheet in an anti(cid:173)
`parallel fashion. The lower part is a rather regular anti-par(cid:173)
`allel /3-pleated sheet. An analysis of the conformational an(cid:173)
`gles of the participating amino acid residues (5- 7, 19-24,
`6~-65, and 71-75) yields mean values for ¢ and it' of -117
`and + 140°, respectively, with standard deviations of 15 and
`16°.
`· The ¢ and it' values for a regular anti-parallel f3 structure
`(f3-poly(L-alanine)) are -139 and + 135° (Arnott et al.,
`1967). The remaining residues of the lower molecular part
`are involved in reverse turns. In the upper part, a regular
`antiparallel /3-structure conformation is formed by two
`strands containing the residues 34-38 and 85-89. The anal(cid:173)
`ysis of the conformational angles yields mean values for ¢
`and .;,- of -116 and + 137° with standard deviations of 13
`and 11°, respectively. The other three strands of that mo(cid:173)
`lecular part are attached in a more irregular (3 structure.
`
`BI 0 CHEMISTRY, V 0 L.
`
`I 4, 1" 0. 2 2,
`
`I 9 7 5 494 7
`
`5 of 10
`
`BI Exhibit 1113
`
`

`

`Table Ill: lntramolecular Hydrogen Bonds.
`
`Main Chain
`
`Thr-5
`NH
`C=O
`Thr-5
`NH
`Scr-7
`C=O
`Ser-7
`Leu-I I NH
`Leu-I I C=O
`Ala-13 C=O
`Ser-14 C=O
`Gln-16 NH
`Ser-17 C=O
`Val-19 C=O
`Val-19 NH
`lle-21
`NH
`lle-21
`C=O
`Cys-23 NH
`Cys-23 C=O
`Ala-25 NH
`Ile-29 C=O
`Ile-29
`NH
`Ile-30 C=O
`Leu-33 C=O
`Asn-34 NH
`Asn-34 C=O
`Trp-35 NH
`Trp-35 C=O
`Tyr-36 NH
`Tyr-36 C=O
`Gln-37 NH
`Gln-37 C=O
`Gln-38 NH
`Gln-38 C=O
`Thr-39 C=O
`Tyr-49 NH
`Tyr-49 C=O
`Arg-61 C=O
`Ser-63 C=O
`Ser-63 NH
`Ser-65 NH
`Ser-65 C=O
`Ala-84 C=O
`Tyr-86 NH
`Tyr-86 C=O
`Cys-88 C=O
`
`Main Chain
`
`Gln-24 C=O
`Gln-24
`NH
`Thr-22 C=O
`Thr-22
`NH
`Lys-103 C=O
`Gln-105 NH
`Thr-107 NH
`Asp-17
`NH
`Leu-78 C=O
`Leu-78
`NH
`!le-75
`NH
`Ile-75
`C=O
`Phe-73 C=O
`Phe-73
`NH
`Tyr-71 C=O
`Tyr-71
`NH
`Thr-69 C=O
`Tyr-32
`NH
`Gly-68 C=O
`Gly-68
`NH
`Ala-51
`NH
`Gln-89 C=O
`Gln-89
`NH
`Ile-48
`C==O
`NH
`Leu-47
`Tyr-87 C=O
`Tyr-87
`NH
`Lys-45 C=O
`Lys-45
`NH
`Thr-85 C=O
`Thr-85
`NH
`Lys-42
`NH
`Asn-53 C=O
`Asn-53
`NH
`Ser-76
`NH
`Thr-74
`NH
`Thr-74 C=O
`Thr-72 C=O
`Thr-72
`NH
`Lcu-104 NH
`Thr- l 02 C==O
`Thr-102 NH
`Gly-99
`NH
`
`Chain 1 Chain 2
`d (A) d (A) <d> (A)
`
`2.9
`2.9
`3.1
`2.7
`3.1
`3.0
`2.5
`3.0
`2.9
`
`3.1
`
`3.0
`2.9
`2.9
`2.9
`2.8
`3.0
`
`3.2
`2.8
`2.8
`2.9
`3.0
`3.0
`2.6
`2.9
`2.9
`2.8
`2.9
`2.9
`
`3.0
`3.0
`2.8
`3.1
`
`3. l
`2.8
`3.0
`2.9
`2.9
`2.8
`
`3.1
`3.0
`3.0
`2.8
`3.0
`3.0
`
`3.1
`2.9
`3.2
`3.0
`3.1
`3.0
`2.8
`2.9
`2.8
`2.9
`2.9
`3.1
`
`2.7
`2.9
`2.9
`2.9
`2.8
`2.6
`2.8
`2.9
`2.8
`2. 7
`3.l
`3.2
`3.0
`2.9
`2.9
`2.9
`2.9
`3.2
`3.0
`3.1
`3.0
`3.1
`2.8
`
`3.0
`3.0
`3.1
`2.8
`3.1
`3.0
`
`3.0
`2.9
`
`3.0
`
`3.0
`2.9
`2.9
`2.9
`2.8
`2.9
`
`2.7
`2.8
`2.9
`3.0
`2.9
`2.6
`2.8
`2.9
`2.8
`2.8
`3.0
`
`3.0
`3.0
`2.9
`3.0
`
`3.1
`2.9
`3. l
`3.0
`3.0
`2.8
`
`Main Chain
`
`Side Chain
`
`Chain 1Chain2
`d(A) d(A) <d>(A)
`
`2.9
`2.8
`2.8
`
`2.7
`2.7
`3.0
`2.6
`
`2.8
`2.7
`2.7
`3.1
`
`2.8
`2.7
`2.9
`
`2.8
`2.7
`2.8
`3.1
`
`- - - -
`C=O
`Thr-97
`Ile-2
`OH
`Thr-102
`C=O
`Pro-8
`OH
`coo(cid:173)
`Asp-17
`:"-IH
`Scr-14
`Ala-25 C=O
`OH
`Thr-69
`coo-
`2.8
`llc-30
`NH
`Asp-28
`coo-
`3.1
`Gln-79 NH
`Asp-82
`2.7
`OH
`Asp-82 C=O
`Tyr-86
`Gln-6
`Tyr-86 C=O
`O=C-NH, 2.7
`Gln-90
`Gln-92 NH
`O==C-NH, 2.9
`Pro-95 C=O
`Gln-90
`O=C-NH 2 3.0
`- - - - -
`Chain !Chain 2
`d(A) d(A) <d)(A)
`
`Side Chain
`
`Side Chain
`
`Gln-6 O=C-NH,
`Asn-34 O=C-NH,
`Tyr-36 OH
`Arg-61 NH 2
`Arg-6 l NH,
`
`Thr-102
`Gln-89
`Gln-89
`Asp-82
`Gln-79
`
`2.9
`OH
`O=C-NH,
`3.1
`O=C-NH,
`3.0
`coo(cid:173)
`2.9
`O=C - NH,
`2.7
`- - - - - - - -
`
`2.9
`
`3.0
`2. 7
`
`2.9
`
`3.0
`2.8
`
`One strand is the carboxyl-terminal end, another the seg(cid:173)
`ment containing the residues 43-50, which makes a kink,
`possibly to allow all three consecutive hydrophobic residues
`(Leu-46, Leu-47, and Ile-48) to lie in the interior of the
`molecule. The third strand is the segment from amino acid
`residues 51-6 l, which forms a large loop containing two re-
`
`EPP ET AL.
`
`__ , ___ __)
`-
`\,____
`FIGURE 3: Hydrogen-bonding scheme of the main-chain atoms of a
`monomer of V REI·
`
`verse turns. In this segment is a deletion of seven amino
`acid residues in the VL part of the Fab' fragment of a
`human myeloma immunoglobulin IgG I(/..) New (Poljak et
`al., 1974 ). Such a deletion must cause a considerable rear(cid:173)
`rangement of the peptide chain, possibly from Tyr-49 on.
`Leu-46 to Ile-48 is conserved in sequence and possibly also
`in position as well as Arg-61. The upper half contains two
`hypervariable regions L2 (amino acid residues 50-56) and
`L3 (89-97), whereas LI (28-34) is an extended polypeptide
`chain connecting both sheets.
`Table III lists the intramolecular hydrogen bonds of the
`two monomers (main-chain bonds, main-chain-side-chain
`bonds and side-chain bonds) and their donor-acceptor dis(cid:173)
`tances. A hydrogen bond was counted when the donor-ac(cid:173)
`ceptor distance was less than 3.2 A and the angles were
`stereochemically reasonable. This is a more stringent condi(cid:173)
`tion than was applied previously, hence less hydrogen bonds
`were found. Figure 3 shows the hydrogen-bonding scheme
`of the main-chain atoms. Most of the hydrogen bonds pos(cid:173)
`tulated in our preliminary model have been confirmed
`(compare Figure 3 with Figure 3(8) in Epp et al., 1974).
`The hairpin bends (Yenkatachalam, 1968; Crawford et al.,
`197 3) could also be confirmed and a few further bends were
`characterized (Table IV). It became clear during the course
`of refinement that two out of six prolines (Pro-8 and Pro-
`95) of the molecule are cis-prolines (Huber and Steige(cid:173)
`mann, 197 4). Both residues lie in the third position of re(cid:173)
`verse turns and are nearly conserved in K-type light chains.
`The presence of these two X-Pro cis-peptide groups in K(cid:173)
`light chains and the observation of cis-Pro bends in several
`other protein molecules suggest that this is an important
`structural element.
`lnterdomain Contacts. The monomer-monomer contact
`can be classified into two parts, the contact region A and
`the hapten binding site B. The amino acid residues Tyr-36,
`Gln-38, Ala-43, Pro-44, Tyr-87, Gln-89, Phe-98, and two
`parts of the polypeptide chain (Lys-42, Gly-99) of both mo(cid:173)
`nomers participate in the contact region. The hydrophobic
`core formed of residues Tyr-36, Ala-43, Pro-44, Tyr-87,
`and Phe-98 encloses a small cavity and is terminated on the
`backside by hydrogen-bonded Gln-38 residues. In the front
`are the polar residues Gln-89 making hydrogen bonds to the
`phenolic groups of both Tyr-36 residues. The arrangement
`of these residues has a twofold symmetry within the limits
`of error; this symmetry is shown in Figure 5. Phe-98 is in(cid:173)
`variant in all light chains: the other residues are almost
`completely conserved.
`The hapten binding site lies in front of the contact region
`along the local twofold axis. Residues Tyr-32, Tyr-36, Tyr-
`49, Tyr-91, Tyr-96, Asn-34, Gln-89, Leu-46, and Leu-94
`from both monomers contribute to the hapten binding site
`forming a slit-like pocket. The spatial arrangement of these
`residues can be seen in Figure 6 viewed along the local diad.
`
`4948
`
`B 1 0 CH EM 1 ST RY. V 0 L.
`
`l 4, N 0. 2 2,
`
`l 9 7 5
`
`6 of 10
`
`BI Exhibit 1113
`
`

`

`REFINED S